Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Positive AI: Key Challenges in Designing Artificial Intelligence for Wellbeing

Willem van der Maden, Derek Lomas, Malak Sadek, Paul Hekkert

TL;DR

The paper advocates Positive AI, arguing that wellbeing should guide AI design and deployment. It presents a cybernetic, sociotechnical framework to analyze four core challenges: modeling, assessing, designing, and optimizing wellbeing. It discusses how to translate qualitative wellbeing experiences into measurable system metrics and how to design interventions that affect wellbeing across platforms. Its contribution lies in combining human-centered design with ethical AI and alignment literatures to articulate an actionable roadmap for creating AI that fosters flourishing while acknowledging pace, tradeoffs, and sociotechnical complexity. Practically, it urges ongoing stakeholder participation, transparency, and reflexive iteration to align AI with long-term human wellbeing.

Abstract

Artificial Intelligence (AI) is a double-edged sword: on one hand, AI promises to provide great advances that could benefit humanity, but on the other hand, AI poses substantial (even existential) risks. With advancements happening daily, many people are increasingly worried about AI's impact on their lives. To ensure AI progresses beneficially, some researchers have proposed "wellbeing" as a key objective to govern AI. This article addresses key challenges in designing AI for wellbeing. We group these challenges into issues of modeling wellbeing in context, assessing wellbeing in context, designing interventions to improve wellbeing, and maintaining AI alignment with wellbeing over time. The identification of these challenges provides a scope for efforts to help ensure that AI developments are aligned with human wellbeing.

Positive AI: Key Challenges in Designing Artificial Intelligence for Wellbeing

TL;DR

The paper advocates Positive AI, arguing that wellbeing should guide AI design and deployment. It presents a cybernetic, sociotechnical framework to analyze four core challenges: modeling, assessing, designing, and optimizing wellbeing. It discusses how to translate qualitative wellbeing experiences into measurable system metrics and how to design interventions that affect wellbeing across platforms. Its contribution lies in combining human-centered design with ethical AI and alignment literatures to articulate an actionable roadmap for creating AI that fosters flourishing while acknowledging pace, tradeoffs, and sociotechnical complexity. Practically, it urges ongoing stakeholder participation, transparency, and reflexive iteration to align AI with long-term human wellbeing.

Abstract

Artificial Intelligence (AI) is a double-edged sword: on one hand, AI promises to provide great advances that could benefit humanity, but on the other hand, AI poses substantial (even existential) risks. With advancements happening daily, many people are increasingly worried about AI's impact on their lives. To ensure AI progresses beneficially, some researchers have proposed "wellbeing" as a key objective to govern AI. This article addresses key challenges in designing AI for wellbeing. We group these challenges into issues of modeling wellbeing in context, assessing wellbeing in context, designing interventions to improve wellbeing, and maintaining AI alignment with wellbeing over time. The identification of these challenges provides a scope for efforts to help ensure that AI developments are aligned with human wellbeing.
Paper Structure (22 sections, 2 figures, 1 table)

This paper contains 22 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Ethical AI
  3. AI Alignment
  4. Why human-centered design?
  5. Why wellbeing?
  6. Framing the challenges: human-centered design of AI systems
  7. Key Challenges
  8. Challenges related to modeling wellbeing
  9. Identifying relevant wellbeing dimensions
  10. Modeling and attributing wellbeing changes
  11. Challenges related to assessing wellbeing
  12. Contextually operationalizing wellbeing
  13. Translating qualitative experiences into system metrics
  14. Challenge related to designing for wellbeing
  15. Methodological challenge of designing (AI) actions to promote wellbeing
  16. ...and 7 more sections

Figures (2)

  • Figure 1: Shows a schematic representation of a cybernetic system. The different challenges can be mapped onto this framework: (1) understanding the system context which entails modeling the relation between wellbeing of the systems constituents and its various components; (2) operationalizing said model of wellbeing; (3) designing interventions to actively promote operationalized model of wellbeing; and (4) retaining alignment with the overall goal. The latter refers to both challenges of algorithmic optimization as well as scrutinizing the objective (e.g., is the wellbeing objective still aligned to needs and desires of all relevant stakeholders?)
  • Figure 2: Different systems operate at different paces, adapted from stephen_pace_2021